Monoamine oxidase (MAO-N) catalyzed deracemization of tetrahydro-β- carbolines: Substrate dependent switch in enantioselectivity

Article abstract of DOI:10.1021/cs400724g

The tetrahydro-β-carboline (THBC) ring system is an important structural motif found in a large number of bioactive alkaloid natural products. Herein we report a broadly applicable method for the synthesis of enantiomerically pure β-carbolines via a deracemization procedure employing the D9 and D11 variants of monoamine oxidase from Aspergillus niger (MAO-N) in combination with a nonselective chemical reducing agent. Biotransformations were performed on a preparative scale, leading to the synthesis of optically enriched products in excellent enantiomeric excess (e.e.; up to 99%) and isolated yield (up to 93%). Interestingly, a switch in enantioselectivity associated with the MAO-N variants is observed as the nature of the C-1 substituent of the THBC is varied. Molecular modeling provided an explanation for this observation and highlighted key active site residues which were modified, resulting in an increase in (R)-selectivity associated with the enzyme. These results provide insight into the factors which influence the selectivity of the MAO-N variants, and may offer a platform for future directed evolution projects aimed toward the challenge of engineering (R)-selective amine oxidase biocatalysts.

Full text of DOI:10.1021/cs400724g

Research Article
pubs.acs.org/acscatalysis
Monoamine Oxidase (MAO-N) Catalyzed Deracemization of
Tetrahydro-β-carbolines: Substrate Dependent Switch in
Enantioselectivity
Diego Ghislieri, Deborah Houghton, Anthony P. Green, Simon C. Willies, and Nicholas J. Turner*
School of Chemistry, Manchester Institute of Biotechnology, University of Manchester, 131 Princess Street, Manchester, M1 7DN,
U.K.
S
* Supporting Information
ABSTRACT: The tetrahydro-β-carboline (THBC) ring system is
an important structural motif found in a large number of bioactive
alkaloid natural products. Herein we report a broadly applicable
method for the synthesis of enantiomerically pure β-carbolines via a
deracemization procedure employing the D9 and D11 variants of
monoamine oxidase from Aspergillus niger (MAO-N) in combina-
tion with a nonselective chemical reducing agent. Biotransforma-
tions were performed on a preparative scale, leading to the synthesis
of optically enriched products in excellent enantiomeric excess (e.e.;
up to 99%) and isolated yield (up to 93%). Interestingly, a switch in
enantioselectivity associated with the MAO-N variants is observed as the nature of the C-1 substituent of the THBC is varied.
Molecular modeling provided an explanation for this observation and highlighted key active site residues which were modified,
resulting in an increase in (R)-selectivity associated with the enzyme. These results provide insight into the factors which
influence the selectivity of the MAO-N variants, and may offer a platform for future directed evolution projects aimed toward the
challenge of engineering (R)-selective amine oxidase biocatalysts.
KEYWORDS: β-carbolines, biocatalysis, chiral amine, monoamine oxidase, deracemization
synthetic community, and numerous pioneering studies
describing the total synthesis of these structures have been
reported.1−4 The development of methods for the asymmetric
synthesis of THBCs continues to represent an important
synthetic challenge and although a number of approaches have
been described previously, including the use of chiral
auxiliaries,5,6 transition metal catalysts,7−9 organocatalysts,10,11
and more recently biocatalysts,12−14 the development of
broadly applicable, alternative strategies remains of great
interest.
INTRODUCTION
■
Tetrahydro-β-carbolines (THBCs) are an important family of
bioactive alkaloids that are widespread in nature. Important
examples include the antileishmanii compound harmicine 1, the
antihypertensive drugs reserpine 2 and ajmalicine 3, and the
stimulant yohimbine 4 (Figure 1). The complex molecular
architectures and the wealth of biological activities associated
with the THBCs have attracted significant attention from the
We have previously described a chemoenzymatic method for
the deracemization of chiral amines to generate the
enantiomerically pure isomer via a redox catalytic cycle (Figure
2).15 Variants of monoamine oxidase from Aspergillus niger
(MAO-N) have been developed in our laboratory, via directed
evolution approaches, which catalyze the enantioselective
oxidation of chiral amines with broad structural features.
When coupled with a nonselective chemical reducing agent
(e.g., ammonia-borane), these engineered enzymes have been
shown to mediate the deracemization of chiral primary,
secondary, and tertiary amines.15−17
Received: August 23, 2013
Revised: October 8, 2013
Figure 1. Representative examples of alkaloids containing the β-
carboline skeleton.
© XXXX American Chemical Society
2869
dx.doi.org/10.1021/cs400724g | ACS Catal. 2013, 3, 2869−2872

ACS Catalysis
Research Article
Figure 2. Generic catalytic cycle for the deracemization of chiral
amines using a combination of an enantioselective amine oxidase with
a nonselective chemical reducing agent.
Recently we reported the application of the D9 variant of
MAO-N in the deracemization of eleagnine, leptaflorine, and
harmicine, three alkaloid natural products belonging to the β-
carboline subclass.18 In this study the D9 variant was found to
be selective for oxidation of the (S)-enantiomers of these
substrates. We now describe the application of MAO-N variants
for the enantioselective oxidation of a series of 1-substituted β-
carbolines, thus providing a broadly applicable strategy for the
synthesis of these structures in optically pure form.
Significantly, a switch in enantioselectivity is observed as the
nature of the C1 substituent is varied. The origins of this
inversion in selectivity are explored, and the results provide a
previously unreported insight into the factors which influence
the enantioselectivity associated with the MAO-N variants.
A panel of aliphatic and aromatic 1-substituted racemic
THBCs was synthesized from tryptamine hydrochloride and
the required aldehyde via a Pictet−Spengler reaction according
to a previously reported procedure13 (see Supporting
Information). The racemic β-carbolines were then subjected
to a deracemization protocol using Escherichia coli whole cells
expressing two different MAO-N variants (D9 and D11) to
compare the activity and enantioselectivity (Figure 3). The
MAO-N D11 variant was recently developed for the oxidation
of structures containing the benzhydrylamine template, and
differs from the D9 variant by a single point mutation (H430G)
which results in an increase in the volume of the small pocket
within the active site.18 Biotransformations were carried out at
15 mM concentration in 1 M phosphate buffer (pH = 7.8)
using 4 equiv of BH3−NH3 as the reducing agent.
Figure 3. MAO-N (D9 and D11) mediated deracemization of β-
carbolines 5a−m. Biotransformation conditions: [S] = 15 mM; [BH3−
NH3] = 60 mM; wet cells =200 mg/mL; 1 M potassium phosphate
buffer (pH 7,8); 37 °C; 250 rpm; pH = 7.8. Reactions were carried out
on a 4.5 μmol scale, e.e. values were determined by chiral HPLC.
The results are shown in Figure 3 and demonstrate that β-
carbolines 5a and 5d−k can be deracemized using either the
D9 or D11 variants (or both), providing the optically enriched
products in good to excellent enantiomeric excess (e.e.). The
racemic 1-methyl substituted THBC 5a was converted to the
corresponding (R)-enantiomer in >99% e.e. with both the D9
and the D11 variant. Replacing the methyl group 5a with the
bulkier ethyl 5b and iso-propyl 5c substituents resulted in a
reduction in e.e. of the products, even after prolonged reaction
times (>72 h). Biotransformations performed on the single
enantiomers (S)-5b and (R)-5b in the absence of reducing
agent show that the MAO-N variants mediate the oxidation of
each enantiomer at comparable rates (see Supporting
Information), providing an explanation for the low e.e.
associated with this substrate. Increasing the size and
lipophilicity of the C1 substituent led to an increase in e.e. of
the deracemized products (substrates 5e−5k). The MAO-N
D11 variant is a particularly effective biocatalyst for the
deracemization of those substrates which possess sterically
demanding C1 substituents, while the D9 variant was shown to
be more effective in the deracemization of the cyclopropyl
derivative 5d. Presumably the increase in active site volume
associated with the D11 variant (cf. MAO-N D9) resulting
from the H430G mutation facilitates the binding of the
sterically demanding substrates. Unfortunately, neither of the
MAO-N variants displayed activity toward the benzyl and −CF3
substituted THBCs (5l and 5m respectively). To demonstrate
the applicability of these MAO-N variants in preparative scale
reactions, the deracemization of substrates 5e−g and 5k was
performed on a 0.4 mmol scale, and the enantiomerically
enriched products were isolated in good yields (>85%) and e.e.
(99% for 5e−g, 90% for 5k).
Determination of the absolute configuration of the amines
derived from the deracemization process established that both
D9 and D11 variants exhibited (S)-selectivity for 5a−b and
(R)-selectivity for 5c−k highlighting a switch in enantioprefer-
ence which is maintained in both variants. In an attempt to gain
insight into the molecular basis underlying this change in
enantioselectivity we decided to examine the different binding
modes of the substrates in the MAO-N active site. Substrates
(S)-5a, (S)-5b, (R)-5b, and (R)-5g were docked into the active
site of the MAO-N D11 variant (PDB id: 3zdn)18 (Figure 4)
using Discovery Studio (Accelrys Software Inc.). The (S)-
enantiomers of 5a and 5b adopt a putative binding mode in
2870
dx.doi.org/10.1021/cs400724g | ACS Catal. 2013, 3, 2869−2872

ACS Catalysis
Research Article
Table 1. MAO-N (D10 and D11) Mediated Deracemization
a
of β-Carbolines 5a−5d
entry
time (h)
MAO-N D10 e.e. (%)
MAO-N D11 e.e. (%)
5a
5b
5c
5e
24
48
48
48
96 (R)
49 (S)
94 (S)
99 (S)
99 (R)
12 (R)
22 (S)
>99 (S)
a
Biotransformation conditions: [S] = 15 mM; [BH3−NH3] = 60 mM;
wet cells = 200 mg/mL; 1 M potassium phosphate buffer (pH 7.8); 37
°C; 250 rpm; pH = 7.8. Reactions were carried out on a 4.5 μmol
scale, e.e. values were determined by chiral HPLC.
substituents such as iso-propyl (5c), MAO-N D10 shows much
higher (R)-enantioselectivity (94% e.e.) compared to the D9
and D11 variants which were both (R)-selective but displayed
only low enantioselectivity (40 and 33% e.e. respectively).
Modeling of (R)-5c into the D10 and D11 active sites reveals
that the substrate is positioned closer to the flavin adenine
dinucleotide (FAD) cofactor in MAO-N D10 (THBC-N2−
FAD-N5 distance 3.56 Å) compared with a distance of 3.90 Å
in the MAO-N D11, thus providing an explanation for the
observed enhanced (R)-selectivity associated with the D10
variant (Figure 5). These results demonstrate that residues in
Figure 4. Docking of (S)-5a (A), (R)-5g (B), (S)-5b (C), and (R)-5b
(D) into the active site of MAO-N D11, highlighting the alternative
binding modes associated with the (R)- and (S)-enantiomers.
which the C1 substituent is buried within the active site and
points away from the entrance channel. Attempts to dock (S)-
5g failed to provide a reasonable productive binding for this
substrate, presumably because of unfavorable steric interactions
between the bulky C1 substituent and active site residues. This
situation is consistent with the experimental observations which
show that the selectivity of the MAO-N variants toward the
(S)-enantiomers decreases as a function of increasing substrate
size. The docking results show that the (R)-enantiomers of 5b
and 5g adopt an alternative binding mode in which the C1
substituent points toward the active site entrance channel.
However, closer inspection of the docked structures suggests
that the THBC is not positioned sufficiently far into the active
site to represent a productive binding mode because of steric
clashes in the region of Ala429. We propose that an increase in
the steric interactions between the C1 substituent and the
amino acid residues around the entrance channel would
effectively force the THBC structure further into the active
site, thus enhancing the probability of accessing a catalytically
productive binding mode. This provides a plausible explanation
for the enhanced (R)-selectivity associated with the MAO-N
variants as the size of the C1 substituent is increased.
Alternatively, modification of the amino-acid residues around
the active-site entrance channel to increase the steric
interactions with the C1 THBC substituents should lead to
an enhanced selectivity toward the (R)-enantiomers and would
provide further evidence in favor of our proposed explanation.
Biotransformations were conducted with the MAO-N D10
variant18 which differs from MAO-N D9 and MAO-N D11 by
four point mutations of residues in the active site entrance
channel (L210F, T213L, Q242M, and T246M). These
modifications have the effect of increasing steric obstruction
around the active site entrance channel. With the MAO-N D10
variant, the switch in enantioselectivity was found to occur with
a substituent as small as an ethyl group (5b), showing relatively
good selectivity toward the (R)-enantiomer, whereas both D9
and D11 showed poor enantioselectivity with a preference for
the (S)-enantiomer (Table 1). In addition, with slightly larger
Figure 5. Docking of (R)-5c in the active site of MAO-N D10 (A) and
MAO-N D11 (B). The THBC-N2−FAD-N5 distance is reduced in
the D10 variant (cf MAO-N D11) as a result of increased steric
obstruction around the active site channel.
the active site channel have an important effect upon the
enantioselectivity of MAO-N variants toward THBC substrates.
It is anticipated that further engineering of these and other
residues around the channel could serve to further optimize the
selectivity of MAO-N toward THBCs and alternative classes of
substrates.
In conclusion, we have shown that chemo-enzymatic
deracemization reactions, using a combination of monoamine
oxidase from A. niger (MAO-N) and a nonselective chemical
reducing agent, can be successfully applied to a range of
different 1-substituted THBCs 5a−k. Interestingly, a switch in
enantioselectivity is observed as the nature of the C-1
substituent is varied. The results of extensive substrate
screening and docking simulations provide valuable insights
into the factors which influence the selectivity of MAO-N
variants, and offer a platform for future directed evolution
projects aimed toward the significant challenge of engineering
enantio-complementary amine oxidase enzymes.
EXPERIMENTAL SECTION
MAO-N Mediated Preparative Deracemization: General
Procedure. In a 50 mL Falcon tube, the required amine
■
2871
dx.doi.org/10.1021/cs400724g | ACS Catal. 2013, 3, 2869−2872

ACS Catalysis
Research Article
hydrochloride salt (0.4 mmol) and BH3−NH3 (1.6 mmol) were
dissolved in potassium phosphate buffer (25 mL, 1 M, pH = 7.8). The
pH of the solution was adjusted to 7.8 by addition of NaOH. Cell
pellet from E. coli cultures (4 g) containing MAO-N D11 was added to
the solution. The tube was placed in a shaking incubator and shaken at
37 °C and 250 rpm. Reactions were monitored by chiral phase HPLC.
Upon completion of the reaction (typically after 48 h), aqueous
NaOH (200 μL, 10 M) and CH2Cl2 (25 mL) were added. The layers
were separated by centrifugation (4000 rpm, 5 min.), and the aqueous
phase was further extracted with CH2Cl2 (20 mL). The combined
organic phases were dried over MgSO4 and concentrated under
reduced pressure.
(17) Dunsmore, C. J.; Carr, R.; Fleming, T.; Turner, N. J. J. Am.
Chem. Soc. 2006, 128, 2224.
(18) Ghislieri, D.; Green, A. P.; Pontini, M.; Willies, S. C.; Rowles, I.;
Frank, A.; Grogan, G.; Turner, N. J. J. Am. Chem. Soc. 2013, 135,
10863.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental procedures and characterization data. This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: Nicholas.turner@manchester.ac.uk.
■
Author Contributions
The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was generously supported by a Marie Curie ITN
(Biotrains FP7-ITN-238531) and a Royal Society Wolfson
Merit Award to N.J.T.
REFERENCES
■
(1) Woodward, R. B.; Bader, F. E.; Bickel, H.; Frey, A. J.; Kierstead,
R. W. J. Am. Chem. Soc. 1956, 78, 2023.
(2) Martin, S. F.; Rueger, H.; Williamson, S. A.; Grzejszczak, S. J. Am.
̈
Chem. Soc. 1987, 109, 6124.
(3) Logers, M.; Overman, L. E.; Welmaker, G. S. J. Am. Chem. Soc.
̈
1995, 117, 9139.
(4) Mergott, D. J.; Zuend, S. J.; Jacobsen, E. N. Org. Lett. 2008, 10,
745.
(5) Gremmen, C.; Willemse, B.; Wanner, M. J.; Koomen, G. J. Org.
Lett. 2000, 2, 1955.
(6) Chen, X. C.; Zhu, J. P. Angew. Chem., Int. Ed. 2007, 46, 3962.
(7) Li, C.; Xiao, J. J. Am. Chem. Soc. 2008, 130, 13208.
(8) Roszkowski, P.; Wojtasiewicz, K.; Leniewski, A.; Maurin, J. K.;
Lis, T.; Czarnocki, Z. J. Mol. Catal. A: Chem. 2005, 232, 143.
(9) Wu, J. S.; Wang, F.; Ma, Y. P.; Cui, X. C.; Cun, L. F.; Zhu, J.;
Deng, J. G.; Yu, B. L. Chem. Commun. 2006, 1766.
(10) Taylor, M. S.; Jacobsen, E. N. J. Am. Chem. Soc. 2004, 126,
10558.
(11) Seayad, J.; Seayad, A. M.; List, B. J. Am. Chem. Soc. 2006, 126,
1086.
(12) Espinoza-Moraga, M.; Petta, T.; Vasquez-Vasquez, M.; Felipe
Laurie, V.; Moraes, L. A. B.; Silva Santos, L. Tetrahedron: Asymmetry
2010, 21, 1988.
(13) Bernhardt, P.; Usera, A. R.; O’Connor, S. E. Tetrahedron Lett.
2010, 51, 4400.
(14) Leipold, F.; Hussain, S.; Ghislieri, D.; Turner, N. J.
ChemCatChem 2013, DOI: 10.1002/cctc.201300539.
(15) Alexeeva, M.; Enright, A.; Dawson, M. J.; Mahmoudian, M.;
Turner, N. J. Angew. Chem., Int. Ed. 2002, 41, 3177.
(16) Carr, R.; Alexeeva, M.; Enright, A.; Eve, T. S. C.; Dawson, M. J.;
Turner, N. J. Angew. Chem., Int. Ed. 2003, 42, 4807.
2872
dx.doi.org/10.1021/cs400724g | ACS Catal. 2013, 3, 2869−2872